Tunnelling and Underground Space Technology 80 (2018) 26–37

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

journal homepage: www.elsevier.com/locate/tust

Innovative solutions for intersection of TBM tunnel and station in Riyadh T

Line 5 metro
⁎
Gonzalo Ramos Schneidera, , Angel Garcia-Fontanet Molinab, Victor Torralba Mendiolac,
Javier Ainchil Lavind
a
Department of Civil and Environmental Engineering, Universitat Politècnica de Catalunya – BarcelonaTech (UPC), Jordi Girona 1-3, 08034 Barcelona, Spain
b
PROGEO Geotechnical Consultants SL, Plà de l’Estany 22, 08192 Sant Quirze del Vallès, Spain
c
Tunnel and Bridge Technologies SL, Fermin Caballero 59, 9C, 28034 Madrid, Spain
d
FAST Consortium, FCC Construccion, Avenida del Camino de Santiago 40, 28050 Madrid, Spain

A R T I C LE I N FO A B S T R A C T

Keywords: One of the main problems when constructing a metro line using a tunnel boring machine (TBM) is the co-
TBM tunnel ordination of the construction of the tunnel with the excavation of the stations that it will intersect, being
Station construction possible to reach the stations before their complete excavation. This paper presents a solution to this problem,
Precast lining one which has been implemented in the Riyadh Metro L5. This solution allows the construction of the tunnel and
the later excavation of the station, while maintaining the operation of the TBM and the tunnel lining.

1. Introduction
Line 5 of the Riyadh Metro (Saudi Arabia) belongs to the package
awarded to the FAST Consortium, of which FCC Construccion is a
partner and responsible for its construction. Other partners in the
Consortium are Samsung C&T, Strukton, Alstom, and local company
Freysinnet. Line 5 consists of a TBM tunnel about 12.9 km long and 12
stations built in open cut. This paper focuses on the problems en-
countered in the intersections between tunnel and stations, as the TBM
reached the stations before they were excavated. For this reason, the
problem of attempting to simultaneously work in the tunnel and, con-
sequently, keeping the lining, while excavating the stations and be-
ginning the erection of their interior structure arose.
This paper presents the problem, the adopted solutions and the
analyses developed for their complete design.
excavation is completed before the tunnel is drilled or, on the contrary,
the excavation of the station is stopped until the tunnel is finished.
Some researchers have studied a similar problem, as Li et al. (2016),
but not for precast segmental linings.
In order to simultaneously excavate the stations and keep working
inside the tunnel, complete structural analyses of the segmental lining
had to be performed. Analysis of precast linings can be developed using
the Elastic Equation Method (ACI, 2014), Beam-Spring Method (Do
et al., 2013; Teachavorasinskun and Chu-upparkan, 2010) or FEM or
DEM simulations (Arnau and Molins, 2015). In this case, a beam-spring
approach has been used because it is precise enough to allow many
calculations, while changing the position of the joints and with different
load combinations.
3. Project and planning

2. Literature review
The construction of underground stations combined with the con-
struction of the tunnel itself has always been a problem. Several station
construction procedures exist, such as cut and cover with different
variations (Liu et al., 2000), NATM with different approaches (Fang
et al., 2012), CAPS (Sadaghiani and Dadizadeh, 2010), BPA (Liu et al.,
2015) and others. Nevertheless, to combine station and tunnel con-
struction is normally something to be avoided. Typically, the station
Line 5 runs southwards from its Depot located in King Abdullah
Road close to the Ministry of Education until the Nations Park in
Central Riyadh. Through its 13 km length it includes 12 underground
stations, running mostly under King Abdulaziz Street which is one of
the main avenues of the capital pf the Kingdom (see Fig. 1). Works
started in October 2013 and all major infrastructures are currently
finished.

⁎
Corresponding author.
E-mail addresses: Gonzalo.ramos@upc.edu (G. Ramos Schneider), agfm@progeo-cga.com (A. Garcia-Fontanet Molina), victor.torralba@btechno.es (V. Torralba Mendiola),
jainchil@fcc.es (J. Ainchil Lavin).

https://doi.org/10.1016/j.tust.2018.05.024
Received 31 March 2017; Received in revised form 10 May 2018; Accepted 28 May 2018
0886-7798/ © 2018 Elsevier Ltd. All rights reserved.
G. Ramos Schneider et al. Tunnelling and Underground Space Technology 80 (2018) 26–37

Fig. 1. Layout of Riyadh Line 5.

Table 1 In general terms, rock parameters are defined according to Hoek

Rock classification. et al. (2002) geomechanical model. Other geotechnical parameters
Rock type Symbol RMR interval RQD interval Weathering
(rock matrix uniaxial compressive strength, rock mass elasticity mod-
grade ulus, etc.) depend on each station location.

Highly weathered HWR 25 ≤ RMR < 40 RQD < 50 3–4 5. Proposed construction method
rock
Moderately MWR 40 ≤ RMR < 55 50 ≤ RQD < 75 2
weathered At stations 5A2, 5A3, 5A5, 5A6, 5B1, 5B2 of the Riyadh metro L5,
rock the TBM advance velocity caused it to reach the stations when they had
Sound rock SR RMR ≥ 55 RQD ≥ 75 0–1 not yet completed their excavation due to several reasons. Initial stu-
dies considered an average progress of 350 m per month, which is
standard for this ground conditions. However, traffic detours and as-
4. Geological and geotechnical summary
sociated utilities diversions were not expedited on time due to the
complex and congested environment and despite the outstanding co-
In general terms, Riyadh geology is defined by a karstic environ-
operation among all the stakeholders. Consequently, it was envisaged
ment with local presence of sinkholes and cavities. Rock mass is pretty
that most of the stations would not be excavated by the time the TBM
massive with an almost horizontal layering. In addition, rock mass
reached them.
weathering degree and presence of discontinuities decrease with depth.
There were two alternatives:
The geological units present in Line 5 area are the following:
Jurassic deposits, Cretacious deposits and Quaternary deposits.
(a) To stop the tunnel boring machine until the excavation of these
The Jurassic deposits are formed by the Jubaila unit (usually de-
stations was complete, to be able to erect the internal structure of
fined in the project as JJ) and the Arab formation, which is in contact
the station before the tunnel crossed them. This alternative had
with the cretaceous units. The later formation has two units: Breccia
very high economic and term costs that made it unfeasible.
complex (J1a) and Carbonate facies (J1c).
(b) To construct the tunnel through the stations, which according to the
The Cretacious deposits are composed by three formations from
project conditions prevented the excavation of the stations to con-
base to top: Suliay formation (C2), Yamama and Buwaib formations
tinue, as the minimum overburden was fixed as 3 m. Obviously, this
(C1). The Suliay formation is composed of compacted micritic to
solution also had economic and time costs.
sparitic limestone with calcarenite and coquina beds.
The Quaternary deposits are composed of the following soil for-
As a result, the technical challenge was to continue the excavation
mations: Sheet Gravel (Qgy) and Wadi Alluvium (Qtz, Qty). The first one
of the stations while maintaining the tunnel operating under con-
is composed of unconsolidated superficial deposits of gravel, sands and
struction. Fig. 2 shows the constructive procedure in which the ex-
silts. The Wadi Alluvium deposits are well-sorted sands.
ecution of the tunnel is compatible with that of the station.
Finally, other Quaternary deposits are the karstic fill materials
The construction procedure, as shown in Fig. 2, included:
(Q1c). They originate from the dissolution processes of limestone,
leaving reddish to brown clays and silt deposits with limestone clast
(a) Excavation up to the tunnel crown (Phases 1 and 2).
that fill karstic voids at depth.
(b) Excavation around the tunnel and stabilization of the lining (Phase
From a geomechanical point of view, the rock units can be classified
3)
as Highly Weathered Rock (HWR), Moderately Weathered Rock (MWR)
(c) Construction of the outer parts of the station countervault while the
and Sound Rock (SR). Classification is carried out according to the
tunnel works continue inside the lining (Phase 3)
following Table 1:
(d) Erection of the lateral parts of the station structure (Phases 4, 5 and

27
G. Ramos Schneider et al. Tunnelling and Underground Space Technology 80 (2018) 26–37

Fig. 2. Construction phases of the tunnel and stations.

28
G. Ramos Schneider et al.
Fig. 3. Ring positions analyzed. The shaded segment is the key.
6)
(e) When the tunnel works finish and all the installations inside the
lining are removed, the lining is dismantled (Phase 7).
(f) Construction of the central part of the countervault (Phase 7)
(g) Erection of the central part of the station structure using precast
elements (Phase 8)
As is well known, lining works by shape thanks to the confinement
provided by the surrounding ground. Without that confinement, the
lining is structurally a mechanism, and would collapse. Consequently, it
was essential to make the ring stable in a provisional and evolutionary
way and without interfering with all the installations and internal
traffic inherent to the execution of the tunnel.
The tunnel is formed by 7 segment rings, with inner diameter of
8.70 m, thickness of 350 mm and an average length of ring of 1.60 m.
There are 19 different ring positions. Reinforced concrete, with a seg-
ment longitudinal reinforcement of 11Φ12 per side (1.60 m segment
width) is used.
Two solutions, a first purely structural one and a second solution
with geotechnical interaction, were adopted to make the tunnel lining
stable while the excavation was progressing around it.
6. Lining stabilization
6.1. First method
In order not to interfere with the works of the tunnel, only the ex-
trados of the lining could be affected. To keep the lining stable, when
the soil was excavated, it was necessary to assure some bending
strength to the longitudinal joints, to avoid collapse. The lining long-
itudinal joints were bolted during construction, and the capacity of the
bolts was used to get that capacity until the joints could be strength-
ened. It is usual to remove the bolts very early after gap injection. In
this case, the bolts were kept in place right up until the lining was
removed.
SAP2000 software was used for structural modelling; analyzing the
structure as evolutionary through the option “nonlinear staged con-
struction analysis”. The model was a single ring, but after an initial
analysis the three positions, 1, 6 and 12 were considered the most
significant, where the key segment was placed at the crown, at the
horizontal diameter and at the lowest possible position (see Fig. 3).
The modulus of lateral and vertical subgrade reaction for a mod-
erately weathered rock are 200,000 kN/m3 and 300,000 kN/m3. In the
analyses the ground is modeled by non-linear springs that operate only
in compression. For the joints, the flexural stiffness corresponding to
29
Tunnelling and Underground Space Technology 80 (2018) 26–37
the joint-bolt assembly obtained from a moment-curvature analysis was
introduced. Given the uncertainties always inherent to ballast coeffi-
cients, a sensitivity analysis of the horizontal module was also per-
formed, which is the one that really influences the problem, adopting
extreme values of 50,000 kN/m3 and of 1,000,000 kN/m3. The springs
are withdrawn as the ground is excavated.
In addition to the self-weight of the lining, a punctual load of
4.45 kN at the crown of the tunnel was included, being the weight of the
ventilation pipe, and a live load corresponding to equipment of 20 kN/
m2 was also considered. Earth-moving equipment (excavator, trucks,…)
can be placed on one side or both sides of the tunnel. An active soil
coefficient of 0.217 was adopted. Also, the load of the tunnel supply
train has been included. In addition, at the final configuration, when
the tunnel and station under construction coexist, an accidental vertical
load of 100 kN was considered at the tunnel crown and a horizontal
load of 100 kN was included at the horizontal diameter. These were
accidental loads which can come from impacts during the construction
of the station lateral structure. All loads have been factored according
to EC2.
Firstly an analysis was done on whether the tunnel could be stable
just considering the bolts at the longitudinal joints. As expected, the
bolts were not able, without any other measure, to stabilize the rings.
Then, the construction sequence for the excavation and stabiliza-
tion, shown in Fig. 4, was developed.
As the tunnel remained in construction, installations and supply
train could not be dismantled. So, all the strengthening operations of
the joints had to be performed from the extrados of the joints. The
simplest way was to use a steel plate, anchored by epoxy fixed steel
bolts. A similar solution with a steel plate for a different problem, but
anchored through an epoxy resin, was used by Zhao et al. (2016). This
solution is optimal if the bending moment acting at the joint produces
tensile forces in the extrados (see Fig. 5).
If the bending moment at the joint produces opening at the intrados
(for example at crown), a solution with a steel plate in the extrados is
useless. The best solution would be a steel plate in the intrados, but the
position of the ventilation pipe blocks its installation. The only way to
act from the extrados is to use a concrete block (Fig. 6). That block was
anchored to both adjacent segments by steel rebars with epoxy. The
block was able to resist the expected bending moments.
As there were 19 different positions of the ring joints, it was es-
sential to determine a simple criterion for each case. That criterion is,
that the joints placed in an arch of 90° centered in the crown should be
strengthened by concrete blocks. The rest of the joints should be
strengthened by steel plates. This criterion was checked for all the
possible solutions (see Fig. 7).
G. Ramos Schneider et al.
A special zone is the transition between tunnel and station. In that
area the tunnel suddenly changes from a section which is confined by
the ground to another which is not, resulting in quite different dis-
placements of the rings. Those differences could cause problems in the
circumferential joints.
In order to study that transition, a finite element 3D analysis was
performed. The lining was modelled by shell elements and boundary
conditions, fixing all the possible displacements, modelling the con-
finement of the soil. Fig. 8 shows the bending moment due to ovali-
zation.
Certainly, there is a distortion of the excavated rings which produce
internal forces at the circumferential joint with the first tunnel ring.
Small Trench to reach the ring crown
Excavation 0.5 m deep over the ring
Concrete
block
Strengthening of existing radial joints from
extrados. Concrete blocks are used.
Fig. 4. Excavation procedure.
30
Tunnelling and Underground Space Technology 80 (2018) 26–37
Nevertheless, maximum bending moments is 110 mkN/m, which can be
resisted with a safety factor of 1.91. The shear forces could be resisted
by the dowels of the segments connecting adjacent rings. The thrust
force of the TBM, which is important for this phenomenon and favor-
able, was neglected in order to be conservative.
Finally a stability analysis of the berm was performed using Plaxis
2D (see Fig. 9).
The solution was used with great success (Figs. 10 and 11). It was
used in stations 5A2, 5B1, 5B2, 5A3 and 5A5 (see Fig. 1).
Excavation of a 0.5 m deep layer
Excavation of a 0.5 m deep layer
Excavation of successive 0.5 m deep layers
up to the tunnel axis.
G. Ramos Schneider et al. Tunnelling and Underground Space Technology 80 (2018) 26–37

Steel
plates

Strengthening of the radial joints from the

Excavation of the berm to permit the
extrados. Steel plates are added
construction of the station
Fig. 4. (continued)

Fig. 5. Strengthening of a radial joint at the extrados by a steel plate.

Fig. 6. Concrete blocks strengthening.

31
G. Ramos Schneider et al. Tunnelling and Underground Space Technology 80 (2018) 26–37

Fig. 7. Criterion for joint strengthening. All joints in Zone A are stregthened by concrete blocks and in Zone B by steel plates.

Fig. 8. Bending moments due to ovalization at the transition zone.

6.2. Second method
The previously described method was a total success. Nevertheless
the strengthening of the joints simultaneously with the excavation was
quite time consuming. Furthermore, during the excavation of some of
the stations it was possible to observe that the quality of the rock was
better than expected. In fact, and according to the field data, most of the
excavation was carried out in Moderately Weathered Rock (MWR) or
Sound Rock (SR). This allowed an improvement in the geotechnical
parameters that resulted in better data from the geomechanical models.
An example of the analyzed data is shown in Fig. 12.
As already explained, the behavior of a ring is adequate when
confined. Following that idea, it was possible to determine the
minimum rock thickness around the tunnel rings in order to keep the
segments confined. If that thickness was low enough, it would be pos-
sible to eliminate all the strengthening of the joints. On the other hand,
the available space will be less than for the first method. Besides, it was
decided to excavate the rock with heavy trenchers that had to cross the
tunnel at the crown.
Geomechanical model geometry is shown in Fig. 13. The following
construction hypotheses have been considered:
• Tunnel ring rotation 0 degrees. Symmetric excavation.
• Tunnel ring rotation 0 degrees. Non-symmetric excavation.
• Tunnel ring rotation 45 degrees. Symmetric excavation.
• Tunnel ring rotation 45 degrees. Non-symmetric excavation.
• Tunnel ring rotation 90 degrees. Symmetric excavation.
• Tunnel ring rotation 90 degrees. Non-symmetric excavation.

32
G. Ramos Schneider et al. Tunnelling and Underground Space Technology 80 (2018) 26–37

Fig. 9. Geomechanical model results. Total displacements contours with the loads of a supply train. Maximum value: 5.22 mm.

Fig. 10. Tunnel crown and lateral excavation with the joints already strenghtened.

Fig. 11. Bottom slab and internal structure of the station is being built.

33
G. Ramos Schneider et al. Tunnelling and Underground Space Technology 80 (2018) 26–37

Fig. 12. Geological Strenght Index (GSI) estimation from field data. Symbols correspond values that arise from site estimations while lines correspond to the GSI
mean value for each geotechnical unit together with a standard deviation confidence interval.

In each case, tunnel lining internal forces, as well as global geo-
technical stability were checked for the above mentioned load combi-
nations. Some examples of the obtained results are shown in Figs. 14
and 15. Fig. 16 shows the station during excavation and the central
tunnel which is surrounded by rock. In an additional calculation it was
determined that a ramp could be kept above the tunnel, being the way
to haul the excavated rock from inside the station. Obtained global
safety factors are in the interval 1.19–4.30, with most of the cases
higher than 2.0. Specific values depend on the problem geometry and
the acting external loads. Minimum values can be regarded as accep-
table for a temporary situation.
This solution was used in station 5A6 (see Fig. 1).
7. Monitoring
No underground structure can be erected without an appropriate
monitoring plan in order to confirm all the hypotheses assumed during
its design. As there was continuous excavation activity outside of the
tunnel, only displacements inside the tunnel could be measured. The
monitoring points are shown in Fig. 17.
Although the most interesting points to measure ovalization are at the
crown and the central diameter, due to the presence of the ventilation, belt
conveyor, pipes, … only the points shown in Fig. 17 were available.
Six monitoring sections where adopted for each station. When the
excavation of the station was more than 6 m over the tunnel crown, 3
readings a week were done. When the excavation was between 3 and

34
G. Ramos Schneider et al. Tunnelling and Underground Space Technology 80 (2018) 26–37

Fig. 13. Geomechanical model geometry. Trencher excavation.

Fig. 14. Tunnel lining ring rotation: 0 degrees. Trencher centered load. Symmetric excavation. Axial forces (tensile forces in blue and positive).

6 m above the tunnel crown, daily measures were obtained. After that,
three readings per day were measured.
The measured displacements were always below the expected ones
(see Table 2 which shows the measurements of just one ring, as an
example), except for one section of one station where the values were
larger than expected. The tunnel lining was moving upward just at one
point. In that station the works were stopped until a local water table,
which was producing the upward pressure, was detected, and the
pressure released by drilling a well. Immediately the tunnel returned to
its original position, behaving as expected.
8. Conclusions
Two successful innovative solutions for making compatible the
construction of a TBM tunnel and a cut and cover station were devel-
oped and used in a real case. The first solution involved strengthening
the radial joints of the rings from the extrados. This solution included
steel plates and concrete blocks. The structural analysis was performed

35
G. Ramos Schneider et al. Tunnelling and Underground Space Technology 80 (2018) 26–37

Fig. 15. Tunnel lining ring rotation: 0 degrees. Trencher eccentric load. Non-symmetric excavation. Failure mechanism geometry.

Fig. 16. Excavation of one station with this solution. Ramp over the tunnel crown.

Table 2
Displacements comparison for ring 508 at the end of the process.
Theoretical Displ. (mm) Measured Displ. (mm)

Monitoring point vertical horizontal vertical horizontal

A −3.82 −6.29 2.30 −0.2

B −3.48 −5.93 1.55 0.65
C −3.48 5.93 1.30 1.45
D 0.0 0.0 1.15 0.0
E 0.0 0.0 1.0 0.7

with a 2D beam-spring model. Interaction with the surrounding rock

was simulated through nonlinear springs acting only in compression.
The solution was used in 5 real stations. Monitoring showed a real
behavior very close to the theoretical one, confirming the accuracy of
Fig. 17. Monitoring points. the beam model with rotational springs.
The second solution used the existing rock to stabilize the ring, but

36
G. Ramos Schneider et al.
Table 3
Comparison of advantages and disadvantages of the proposed solutions.
Advantages
Both Solutions • Important reduction on schedule
• Excavation of stations is not stopped when the tunnel is reached.
Equipment and manpower can keep working
• Demolishing the tunnel is easier as the segments can be disassembled
Solution 1
Solution 2 • No joint strengthening is needed, so construction time and cost are
reduced
losing some free space from the first one. This solution was analyzed by
a geomechanical 2D FEM model and used in one station as the inter-
action between soil and structure is even more essential in these cases.
Again, the monitoring confirmed the numerical modelling.
The advantages and disadvantages of the proposed solutions are
described in Table 3.
These solutions reduced the schedule of the whole line construction
by more than 4 months.
References
ACI, 2014. ACI 544 ETR Design and Construction of fiber-Reinforced Precast Concrete
Tunnel Segments.
Arnau, O., Molins, C., 2015. Theoretical and numerical analysis of the three-dimensional
response of segmental tunnel linings subjected to localized loads. Tunn. Undergr.
Space Technol. 49, 384–399. http://dx.doi.org/10.1016/j.tust.2015.05.012.
Do, N.A., Dias, D., Oeste, P., Djeran-Maigre, I., 2013. 2D numerical investigation of
segmental tunnel lining behavior. Tunn. Undergr. Space Technol. 37, 115–127.
http://dx.doi.org/10.1016/j.tust.2013.03.008.
37
Tunnelling and Underground Space Technology 80 (2018) 26–37
Disadvantages
• Station structure must be adapted to the construction method
• Invert is built in 3 parts, which makes the connection critical
• The central part of the station should be precast
• There is little room to build the station structure. So the productivity is below
the expected for normal construction.
• Excavation around the tunnel must be with light equipment and the lining
joints must be strengthened
• Good rock/soil strength is needed
Fang, Q., Zhang, D., Yuen Wong, L.N., 2012. Shallow tunnelling method (STM) for
subway station construction in soft ground. Tunn. Undergr. Space Technol. 29,
10–30. http://dx.doi.org/10.1016/j.tust.2011.12.007.
Hoek, E., Carranza-Torres, C., Corkum, B., 2002. Hoek-Brown failure criterion – 2002
Edition. In: Proc. NARMS-TAC Conference, Toronto. 1, 267–273.
Liu, W., Luo, F., Mei, J., 2000. A new construction method for a metro station in Beijing.
Tunn. Undergr. Space Technol. 15 (4), 409–413.
Liu, J., Wang, F., He, S., Wang, E., Zhou, H., 2015. Tunn. Undergr. Space Technol. 49,
130–143.
Sadaghiani, M.H., Dadizadeh, S., 2010. Study on the effect of a new construction method
for a large span metro underground station in Tabriz-Iran. Tunn. Undergr. Space
Technol. 25, 63–69. http://dx.doi.org/10.1016/j.tust.2009.08.004.
Teachavorasinskun, S., Chu-upparkan, T., 2010. Influence of segmental joints on tunnel
lining. Tunn. Undergr. Space Technol. 25, 490–494. http://dx.doi.org/10.1016/j.
tust.2010.02.003.
Li, Y., Jin, X., Lv, Z., Dong, J., Guo, J., 2016. Deformation and mechanical characteristics
of tunnel lining in tunnel intersection between subway station tunnel and construc-
tion tunnel. Tunn. Undergr. Space Technol. 56, 22–33. http://dx.doi.org/10.1016/j.
tust.2016.02.016.
Zhao, H., Liu, X., Bao, Y., Yuan, Y., Bai, Y., 2016. Simplified nonlinear simulation of shield
tunnel lining reinforced by epoxy bonded steel plates. Tunn. Undergr. Space Technol.
51, 362–371. http://dx.doi.org/10.1016/j.tust.2015.10.004.